Hand-marked human face images categorized with bounding box labels in YOLO format. The dataset contains filtered visual data specifically prepared for training and validating facial detection models.
Use Cases
- Train a YOLOv8 or similar object detection model using the provided face labels.
- Benchmark facial localization accuracy using the hand-marked ground truth coordinates.
- Develop a preprocessing pipeline for face-cropping based on the YOLO format bounding boxes.
Strengths
- Features hand-marked bounding box annotations for every face image.
- Labels are provided in the .txt YOLO format for direct model ingestion.
- Dataset includes a filtered selection of human face imagery to ensure annotation accuracy.